mihirp1998 / AlignProp

AlignProp uses direct reward backpropogation for the alignment of large-scale text-to-image diffusion models. Our method is 25x more sample and compute efficient than reinforcement learning methods (PPO) for finetuning Stable Diffusion
256Updated 2 months ago

Alternatives and similar repositories for AlignProp:

Users that are interested in AlignProp are comparing it to the libraries listed below